© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 1-RES-SM
    11. 23936
    Prev
    Next
    INDEX
    Explanations

    the term "offensive" or variations of it

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/1-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.1.hook_resid_post
    Hook Layer
    1
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Neuron Alignment
    Index
    Value
    % of L₁
    464
    +0.12
    0.7%
    443
    +0.11
    0.6%
    25
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    25
    +0.12
    0.03
    308
    +0.11
    0.02
    464
    +0.11
    0.02
    Negative Logits
    ogly
    -1.51
    mary
    -1.46
    RCC
    -1.37
     darling
    -1.34
    uria
    -1.34
    ^âĪĴ
    -1.33
    idd
    -1.29
    poons
    -1.27
     shape
    -1.26
    mie
    -1.25
    POSITIVE LOGITS
    loaded
    2.04
    ensa
    1.99
    ermost
    1.95
    ices
    1.89
    enthal
    1.85
    eting
    1.80
    fe
    1.69
    load
    1.65
    spring
    1.65
    shot
    1.59
    Activations Density 0.117%

    No Known Activations